NeuroImage 47 (2009) 342–352

Contents lists available at ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r.com/ locate /yn img

Gender differences in the functional and structural neuroanatomy ofmathematical cognition

Katherine Keller a, Vinod Menon a,b,c,⁎a Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94305, USAb Program in Neuroscience, Stanford University School of Medicine, Stanford, CA 94305, USAc Symbolic Systems Program, Stanford University, Stanford, CA 94305, USA

Abbreviations: DMN, default mode network; fMRI, fimaging; IPS, intra-parietal sulcus; MA, mental arithgyrus; PPC, posterior parietal cortex; ROI, region of inte⁎ Corresponding author. Program in Neuroscience, Sy

Department of Psychiatry and Behavioral Sciences, 780University School of Medicine, Stanford, CA 94305-5778

E-mail address: menon@stanford.edu (V. Menon).

1053-8119/$ – see front matter © 2009 Elsevier Inc. Aldoi:10.1016/j.neuroimage.2009.04.042

a b s t r a c t

a r t i c l e i n f o

Article history:Received 18 June 2008Revised 13 March 2009Accepted 8 April 2009Available online 17 April 2009

Keywords:GenderSex differencesMental arithmeticfMRIFrontal lobeParietal cortexIntra-parietal sulcusLingual gyrusParahippocampal gyrusDorsal visual streamVentral visual streamDefault mode network

Despite ongoing debate about the nature of gender differences in mathematics achievement, little is knownabout gender similarities and differences in mathematical cognition at the neural level. We used fMRI tocompare brain responses in 25 females and 24 males during a mental arithmetic task involving 3-operandaddition and subtraction. We also used voxel-based morphometry (VBM) to examine gender differences inbrain structure. Although females and males did not differ in accuracy or response times (effect size db0.3),significant gender differences in functional brain activation were observed in the right dorsal and ventralvisuospatial information processing streams (dN1.1). Males showed greater dorsal stream activation in theright intra-parietal sulcus areas important for numerical cognition, and angular gyrus regions of the defaultmode network that are typically deactivated during complex cognitive tasks, as well as greater ventral streamactivation in the right lingual and parahippocampal gyri. VBM revealed an opposite pattern of genderdifferences—compared to males, females had greater regional density and greater regional volume in dorsaland ventral stream regions where males showed greater fMRI activation. There were no brain areas wherefemales showed greater functional activation than males, and no brain areas where males showed greaterstructural density or volume than females. Our findings provide evidence for gender differences in thefunctional and structural organization of the right hemisphere brain areas involved in mathematicalcognition. Together with the lack of behavioral differences, our results point to more efficient use of neuralprocessing resources in females.

© 2009 Elsevier Inc. All rights reserved.

In the past few decades, scientists and the public alike havedebated the existence of gender differences in mathematical skills(Armstrong, 1981; Geary, 1996; Halpern et al., 2007; Hyde, 2005;Kegel-Flom and Didion, 1995). The question of gender differences inmathematics has been particularly contentious because of its wide-spread implications for academic achievement, career choice, andprofessional success in engineering and the sciences (Jacobs, 2005).Despite the stereotypic view that males are “better” at mathematics,evidence for gender differences inmathematics has beenmixed; somestudies suggest differences favoring males, and others suggestdifferences favoring females (Gallagher and Kaufman, 2005). Cogni-tive and educational psychologists have advanced a number of reasonsfor potential gender differences in mathematical cognition; however,

unctional magnetic resonancemetic; PHG, parahippocampalrest; RT, reaction time.mbolic Systems Program, andWelch Rd, Room 201, Stanford, USA.

l rights reserved.

the neurobiological bases of these differences, if any, remaincompletely unknown. We used a cognitive neuroscience approach toinvestigate whether brain and cognitive processes underlying basicmathematical problem solving differ between females and males.

The literature on gender differences in mathematical abilities andachievement has been complex and inconsistent (Gallagher andKaufman, 2005; Halpern et al., 2007). A detailed meta-analysis ofgender differences using over 100 studies found that performance ona range of mathematical tasks was similar in females andmales (Hydeet al., 1990). Several other studies, however, have suggested that malesoutperform females in standardized tests of mathematical achieve-ment (Gallagher and Kaufman, 2005). Although the nature and originof gender differences in mathematics remains unclear, variability inperformance tends to be higher in males (Feingold, 1992; Geary, 1998;Willingham and Cole, 1997). Consequently, males tend to outperformfemales in select groups of high achieving and gifted children andcollege students (Benbow et al., 2000; Hyde et al., 1990; Stanley et al.,1994), whereas females tend to outperform males at the lower end ofthe distribution (Barbaresi et al., 2005). Recent studies, however,suggest that females' underperformance in math relative to males is

343K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

eliminated in more gender-equal cultures (Guiso et al., 2008), and inthe United States, where the math gender gap has been closing overtime (Goldin et al., 2006).

Recent brain imaging studies have contributed tremendously toour understanding of the cognitive processes involved in mathema-tical cognition (Dehaene et al., 2004; Delazer et al., 2003;Menon et al.,2000b). Both lesion and functional imaging studies have identified anetwork of brain regions that are important for mathematicalcognition. These regions include the posterior parietal cortex (PPC),inferior frontal cortex, premotor cortex and basal ganglia. Collectively,these studies suggest that PPC areas, notably cortical areas in andaround the intra-parietal sulcus (IPS), play an important role inmathematical calculation and fact retrieval (Dehaene et al., 2004;Dehaene et al., 1999;Menon et al., 2000b; Rivera et al., 2005). Previousstudies in mixed groups of college-aged females and males havesuggested that PPC and prefrontal cortex contributions to MA taskscan be dissociated, and that PPC regions, notably the IPS, play animportant role in calculation independent of other processingdemands (Menon et al., 2000b). All of these studies have usedmixed groups of males and females with relatively small sample sizesthat preclude determination of whether female and male brains differin the way they process basic mathematical information.

In this study, we used functional magnetic resonance imaging(fMRI) to examine gender similarities and differences during amathematical cognition task in young adults. In order to maximizeour ability to examine differences in functional brain organizationindependent of behavioral differences, we chose a population of well-educated college-age males and females for whom themath cognitiontasks we used would be similar in difficulty (Guiso et al., 2008). Weused a 3-operand mental arithmetic (MA) task involving simpleaddition and subtraction operations similar to those used in Menonet al. (2000a,b), where participants were presented with equations ofArabic (e.g. 4+3−1=5) numerals, and asked to judge whether theequation was correct or incorrect. We used three-operand equationsbecause this task is known to generate reliable and robust responses inPPC, prefrontal and subcortical regions involved in mathematicalcognition, while at the same time providing variability in performanceto facilitate examination of gender differences in brain–behaviorrelations. Finally, we examined whether any observed gender dif-ferences in functional activations were related to the underlying ana-tomical structure using voxel-based morphometry (VBM; Ashburnerand Friston, 2000; Good et al., 2001). Our analysis focused on graymatter density and volume after non-linear corrections for brainvolume so that local differences in brain structure and function couldbe compared more directly (Ilg et al., 2008).

Accordingly, our aims were to (1) examine whether males andfemales engage the same brain areas during MA, (2) quantify theextent of similarities in brain response across males and females, (3)investigate whether the two groups engage similar brain processes inrelation to performance, and (4) examine whether gender differencesin function were related to underlying anatomical differences. Basedon other related research on gender differences in cognitive tasksinvolving mental rotation and working memory (Bell et al., 2006;Frings et al., 2006; Goldstein et al., 2005; Thomsen et al., 2000; Weisset al., 2003), we predicted that (1) behavioral performance would besimilar in males and females, (2) gender differences in brainresponses during mathematical cognition would exist even in theabsence of overt gender differences in behavior, (3) males and femaleswould show extensive overlap in activation of the left and right IPSand angular gyrus, regions important for arithmetic calculation andfact retrieval, and (4) males would show greater activation in the PPCwhereas females would show greater responses in the left and rightfrontal cortex. Understanding whether males and females differ in thebrain systems they use for processing mathematical information mayinform us about potential differences in strategy use and underlyingvulnerability to stereotyping. In order to make appropriate links with

prior behavioral and meta-analysis studies, we report effect sizes forgender differences in brain response. To our knowledge, few, if any,brain imaging studies of gender differences in cognitive function havereported effect sizes.

Method

Participants

Forty-nine healthy, right-handed, adults (25 females and 24males) participated in this study. They ranged from 18 to 36 years ofage, with a mean age of 23.99 (SD=4.66). Mean ages of males(M=23.53, SD=4.92) and females (M=24.36, SD=4.51) were notsignificantly different, t(36)=0.542, ns. Participants were recruitedfrom the Stanford University community after givingwritten informedconsent. All protocols were approved by the Human SubjectsCommittee at Stanford University School of Medicine, and participantswere treated in accordance with the APA “Code of Conduct”.

Experimental design

All participants performed an fMRI experiment involving twoconditions—mathematical calculation (Calculation) and numberidentification (Identification). Participants were presented with 16alternating blocks of Calculation and Identification trials, each blocklasting 32 s. Each Calculation block consisted of eight 3-operandequations of the form a+b−c=d (e.g. 5+4−2=7), where a, b, cand d were all single-digit numerals ranging from 1 to 9. Eachequation was presented for 3.5 s followed by a blank screen for 0.5 s.Participants were instructed to indicate whether the equation wascorrect (e.g., 4+5−2=7) or incorrect (e.g., 4+5−2=8). Half ofthe equations presentedwere correct, and the other half incorrect; theorder of correct and incorrect equations was randomized. EachIdentification block consisted of eight 7-symbol strings (e.g.,4@3&2#5), with numbers and symbols alternating in the same wayas stimuli used in the Calculation task. None of the symbols used inthese stimuli were related to mathematical calculation symbols. Eachstring was presented for 3.5 s followed by a blank screen for 0.5 s.Participants were instructed to indicate whether the string containedthe numeral 5. Half of the strings presented contained the numeral 5;the placement of the 5 within the string and the order of presentationof these strings were randomized. The Identification task wasdesigned to control for basic number and visual processing, motorresponse and low-level attention (Menon et al., 2000b). Ourexperimental design therefore allowed us to examine genderdifferences during MA more precisely than would have been possiblewith a purely perceptual baseline condition.

Behavioral data analysis

Mean reaction time (RT) and percentage of correct responses forthe Calculation and Identification trials were computed for eachparticipant. RT and accuracy were analyzed using a two-way analysisof variance (ANOVA) with factors: gender (Males, Females) and taskcondition (Calculation, Identification).

Brain Imaging

fMRI data acquisitionImages were acquired on a 3T GE Signa scanner using a standard

GEwhole head coil (software Lx 8.3). Headmovement was minimizedduring scanning by a comfortable custom-built restraint. A total of 28axial slices (4.5 mm thickness) parallel to the AC-PC line and coveringthe whole brain were imaged using a T2⁎ weighted gradient echospiral pulse sequence (TR=2000 ms, TE=30 ms, flip angle=70°, 1interleave; Glover and Lai, 1998). The field of view was 20 cm, and the

Table 1Conjunction of brain activation in males and females during mental arithmetic.

Condition Region BA Corrected p value # of voxels Peak Z score Peak MNI coordinates (mm)

x y z

Conjunction (males and females)(Calculation–identification)

L intra-parietal sulcus 7 b0.001 12342 7.52 −22 −70 42L superior parietal lobule 7 7.36 −26 −64 46L supramarginal gyrus 40 7.17 −30 −54 44L inferior frontal gyrus 44 b0.001 15151 7.07 −46 12 28L inferior frontal gyrus/anterior insula 47, 48 6.68 −32 24 0R inferior frontal gyrus/anterior insula 47, 48 6.42 32 24 −2R supramarginal gyrus 40 b0.001 4287 6.37 36 −50 44R intra-parietal sulcus 7 6.26 30 −70 46R intra-parietal sulcus 7 6.19 30 −66 38

Overlap in brain activation across males and females during Calculation, compared to the Identification, condition (pb0.01), as assessed using a conjunction analysis. The threemajoractivation peaks in each of the three significant clusters cluster are shownwith their corresponding Brodmann Area (BA), significance level of the cluster, the number of voxels in thecluster, the Z score of the peak activated voxel, and the MNI coordinates.

344 K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

matrix size was 64×64, providing an in-plane spatial resolution of3.125 mm. To reduce blurring and signal loss arising from fieldinhomogeneities, an automated high-order shimming method basedon spiral acquisitions was used before acquiring functional MRI scans(Kim et al., 2002). To aid in the localization of functional data, a highresolution T1 weighted fast spoiled grass gradient recalled (FSPGR)inversion recovery 3D MRI sequence was used with the followingparameters: TI=300 ms, TR=8 ms; TE=3.6 ms; flip angle=15°;22 cm field of view; 124 slices in coronal plane; 256×192 matrix; 2NEX, acquired resolution=1.5×0.9×1.1 mm. The images werereconstructed as a 124×256×256 matrix with a 1.5×0.9×0.9 mmspatial resolution. Structural and functional images were acquired inthe same scan session.

fMRI data analysis

Preprocessing. The first 5 volumes were not analyzed to allow forsignal equilibration effects. A linear shim correction was appliedseparately for each slice during reconstruction using a magnetic fieldmap acquired automatically by the pulse sequence at the beginning ofthe scan (Glover and Lai, 1998). Functional MRI data were thenanalyzed using SPM5 analysis software (http://www.fil.ion.ucl.ac.uk/spm). Images were realigned to correct for motion, corrected forerrors in slice-timing, spatially transformed to standard MNI spaceusing the echo-planar imaging template provided with SPM5,resampled every 2 mm using sinc interpolation and smoothed witha 4 mm full-width half-maximum Gaussian kernel to decrease spatialnoise prior to statistical analysis. Translational movement in milli-meters (x, y, z) and rotational motion in degrees (pitch, roll, yaw) wascalculated based on the SPM5 parameters for motion correction of thefunctional images in each subject. None of the participants hadmovement greater than 3 mm translation or 3° of rotation.

Individual subject analyses. Task-related brain activation wasidentified using a general linear model. Individual subject analyseswere first performed by modeling task-related conditions. For themathematical cognition task, brain activity related to the two taskconditions (Calculation and Identification) was modeled using boxcarfunctions convolved with a canonical hemodynamic response func-tion and a temporal dispersion derivative to account for voxel-wiselatency differences in hemodynamic response. Low-frequency drifts ateach voxel were removed using a high-pass filter (0.5 cycles/min) andserial correlations were accounted for by modeling the fMRI timeseries as a first degree autoregressive process (Friston et al., 1997).Voxel-wise t-statistics maps for each condition were generated foreach participant using analysis of variance, along with the respectivecontrast images.

Group analyses. Group analysis was performed using the contrastimages and a second-level analysis of variance. Voxel-wise t-statisticsmaps were then used to determine group-level activation for eachtask condition. Significant clusters of activation were determinedusing a voxel-wise statistical height threshold of (pb0.01), withcorrections for multiple spatial comparisons at the cluster level(pb0.05). The following two within-group comparisons were per-formed: (1) males (Calculation minus Identification); and (2) females(Calculation minus Identification); two between-group comparisonswere then performed to examine gender differences in activation anddeactivation: (3) males (Calculation minus Identification)–females(Calculation minus Identification); and (4) females (Calculationminus Identification)–males (Calculation minus Identification). Con-junction analysis (Friston et al., 2005) was subsequently performed todetermine brain regions with themost consistent overlap in activationacross males and females. Activation foci were superimposed on highresolution T1-weighted images and their locations were interpretedusing knownneuroanatomical landmarks (Duvernoy et al., 1999). MNIcoordinates were transformed to Talairach coordinates using a non-linear transformation (Brett, 2000).

ROI analyses. Region of Interest (ROI) analyses were used to furtherquantify brain responses. In the between-group comparisons, whichinvolve double subtractions, additional analyses were conducted todistinguish between differences arising from Calculation task-relatedincreases in activation, as opposed to Identification task-relateddecreases in activation (or “deactivation”). For each significant clusterthat was detected in these comparisons, mean Z scores were firstcomputed for males (comparison #1 above) and females (comparison#2). Clusters in comparisons 3 and 4 were then categorized as arisingfrom deactivation if (a) the mean Z score was significantly negative incomparisons #2 and #1, respectively; and (b) the mean Z score wasnot significantly different from 0 in comparisons #1 and #2,respectively. Both the magnitude and sign of the activations wereexamined in order to ascertain whether any observed brain–behaviorrelations were explained by task-related activation or deactivation.ROI analyses were conducted using the MaRsbar toolbox (http://marsbar.sourceforge.net). ROIs were based on clusters that showedsignificant activation and deactivation differences between males andfemales (Table 1). In order to elucidate the nature of genderdifferences arising from deactivation, we examined brain responsesin regions that overlapped with the default mode network (DMN), asdefined in a previous study in our lab (Greicius et al., 2003). We alsoexamined gender differences in 10 mm spheres around the activationpeaks shown in Table 1. Finally, in order to further examine genderdifferences in activation, functional ROIs based on activation clustersfrom our previous study (Menon et al., 2000b), were also used. We

345K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

used these functional ROIs because they have been found to beconsistently activated in several studies of mathematical cognition.

Average signal strength, as measured by average t-score, and thespatial extent of activation, as measured by the percentage of voxelsactivated above a threshold of ZN2.33 or Zb−2.33, were calculatedfor each ROI and compared across genders using an ANOVA.

Brain–behavior relationships. Finally, brain–behavior relationshipswere also examined using ROI analysis. Signal levels in each ROI wereregressed against accuracy and RT separately inmales and females andthe slopes compared using a difference of slopes test. Regressionanalysis was used to examine whether the relation betweenperformance (either accuracy or RT) and brain activation duringCalculation versus Identification differed between males and females.Contrast images from each participant were entered into theregression analysis along with the behavioral covariates of interest.Positive and negative correlations in brain activation with increasedaccuracy or reaction timewere compared inmales and females. Voxel-wise t-statistics maps were generated and the clusters with significantcorrelations were determined using a voxel-wise statistical heightthreshold of (pb0.01), with corrections for multiple spatial compar-isons at the cluster level (pb0.05).

VBM analysisThirty-four participants (17 males, 17 females) had T1-weighted

images that could be used in VBM analysis of gender differences inbrain anatomy. The remaining 15 participants either did not havestructural MRI data or the data was of inadequate quality for VBManalysis. We performed optimized VBM using the VBM5 toolbox inSPM5 (http://dbm.neuro.uni-jena.de/vbm/vbm5-for-spm5/). T1-weighted images from each participant were first segmented intogray, white and CSF images (Ashburner and Friston, 2005). Segmentedgray matter images were modulated using non-linear warpingprocedures that correct for global brain differences and alignhomologous brain regions into a common space (Ashburner andFriston, 2000; Ilg et al., 2008). The resulting images were thensmoothed using a Gaussian kernel with 10 mm FWHM (Ashburnerand Friston, 2000). Between-group analysis was performed usingSPM5 on the these smoothed images to determine brain voxels wherelocal gray matter density and volume differed between males andfemales, after controlling for differences in brain size.

Fig. 1. Mental arithmetic task performance in males and females. (A) Accuracy as a functidifference between genderswas found in either the Calculation or Identification conditions, afemales, accuracy was significantly greater in the Identification, compared to the Calculatidifference between genderswas found in either the Calculation or Identification conditions, afemales, reaction time was significantly greater in the Calculation, compared to the Identifi

Results

Behavior

AccuracyAn Analysis of Variance (ANOVA) was performed on accuracy with

factors task condition (Calculation and Identification) and gender(Males and Females; Fig. 1A). There was a significant main effect oftask condition, such that participants performed more accurately onthe Identification task (M=0.96, SE=0.01) than on the Calculationtask (M=0.91, SE=0.01), F(1,47)=13.36, pb0.01, partial η2=0.22.There was, however, no significant interaction between task conditionand gender, F(1,47)=2.16, p=0.15, partial η2=0.04. Importantly,there was no main effect of gender; both males (M=0.94, SE=0.01)and females (M=0.93, SE=0.01) were equally accurate on both theCalculation and Identification conditions, F(1,47)=0.61, p=0.44,partial η2=0.01, Cohen's d=0.23. Post-hoc t-tests were performedto confirm that there was no difference in accuracy between malesand females on either of the Calculation and Identification tasksindividually. There was no significant difference in accuracy betweenmales (M=0.93, SE=0.02) and females (M=0.90, SE=0.02) onCalculation trials, t(47)=1.22, p=0.23. There was also no significantdifference in accuracy between males (M=0.96, SE=0.07) andfemales (M=0.96, SE=0.05) on Identification trials, t(47)=−0.14,p=0.89.

Reaction timeAn ANOVA was performed on RT with factors task condition

(Calculation and Identification) and gender (Males and Females;Fig. 1B). As expected, there was a significant effect of task conditionsuch that participants performed more quickly on the Identificationtask (M=938.23 ms, SE=31.13) than on the Calculation task(M=2041.65 ms, SE=48.31), F(1,47)=728.22, pb0.01, partialη2=0.94. There was, however, no significant interaction betweentask condition and gender, F(1,47)=0.09, p=0.76, partial η2=0.00.Importantly, there was no main effect of gender, F(1,47)=0.86,p=0.36, partial η2=0.02, Cohen's d=0.28; males (M=1522 ms,SE=50.17) and females (M=1457 ms, SE=49.16) performed equallyquickly. Post-hoc t-tests were performed to confirm that there was nodifference in RT between males and females on either of theCalculation and Identification tasks individually. There was no

on of gender (male, female) and condition (Calculation, Identification). No significantnd therewas no significant interaction between gender and condition. In bothmales andon, condition. (B) Reaction time as a function of gender and condition. No significantnd therewas no significant interaction between gender and condition. In bothmales andcation, condition.

Fig. 2. Conjunction of brain activation in males and females during mental arithmetic.Similarities in brain activation during the Calculation, compared to the Identification,task in males and females (pb0.01), as determined using conjunction analysis.Consistent activation across the two groups was detected bilaterally in the inferiorfrontal gyrus and adjoining anterior insula, anterior cingulate cortex, pre-supplemen-tary motor area, intra-parietal sulcus and adjoining supramarginal gyrus, lateraloccipital cortex, caudate nucleus and the cerebellum.

Fig. 3. Gender differences in brain activation during mental arithmetic. Significantgender differences in activation were observed in two clusters encompassing (A) thedorsal visual stream and (B) the ventral visual stream during the Calculation, comparedto the Identification, condition (pb0.01, corrected for multiple spatial comparisons). Inthe dorsal stream, males showed greater activation in the right intra-parietal sulcus(IPS) and the right angular gyrus. In the ventral stream,males showed greater activationin the right parahippocampal gyrus and right lingual gyrus. All gender differences werelateralized to the right hemisphere. No brain areas showed greater activation in females,compared to males.

346 K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

significant difference in RT betweenmales (M=2080.37, SE=60.55) andfemales (M=2002.94, SE=74.71) on Calculation trials, t(47)=0.80,p=0.43. There was also no significant difference in RT between males(M=964.47, SE=44.86) and females (M=911.99, SE=43.19) onIdentification trials, t(47)=0.84, p=0.40.

Brain imaging

Brain areas that showed an overlap in activation across males andfemales during MA

Conjunction analysis indicated that both males and females hadsignificant overlap in the Calculation compared to Identification tasks

bilaterally in the inferior frontal cortex (BA 44, 47) and adjoininganterior insula (BA 48), anterior cingulate cortex (BA 24), pre-supplementary motor area (BA 6), intra-parietal sulcus (IPS) andadjoining superior parietal lobule (BA 7) and supramarginal gyrus (BA40), lateral occipital cortex (BA 19), caudate nucleus and cerebellum(Fig. 2, Table 1).

Gender differences in brain activation during MAA between-group comparison revealed significantly greater acti-

vation in males, compared to females, in two right hemisphereclusters, encompassing the dorsal (Cohen's d=1.12) and ventral(Cohen's d=1.19) visual streams, respectively (Fig. 3, Table 2). Thedorsal stream cluster consisted of the right IPS (BA 7) and the rightangular gyrus (BA 39). The ventral stream cluster included the rightlingual gyrus (BA 18) and the right parahippocampal gyrus (PHG; BA37). No brain regions showed greater activation in females comparedto males.

Table 2Gender differences in brain activation during mental arithmetic.

Condition Region BA Cluster p value Effect size # of voxels Peak Z score Peak MNI coordinates(mm)

x y z

Males–females (calculation–identification)Dorsal stream R intra-parietal sulcus (IPS) 7 0.049 1.12 306 3.78 32 −78 48

R angular gyrus 39 0.049 2.88 40 −78 42Ventral stream R lingual gyrus 18 0.022 1.19 355 3.58 8 −54 0

R parahippocampal gyrus (PHG) 37 0.022 3.05 8 −40 0Females–males (calculation–identification)No significant clusters

Gender differences were observed in two clusters, one encompassing the dorsal visual stream and the other the ventral visual stream during the Calculation, compared to theIdentification task (pb .01, corrected for multiple spatial comparisons at pb0.05; cluster extent=306 voxels). Localization of brain activation in the two clusters are shown, alongwith the corresponding Brodmann Area (BA), significance level of the cluster, effect size indicated by Cohen's d value, the number of voxels in the cluster, the Z score of the peakactivated voxel, and the MNI coordinates.

347K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

ROI analysis of gender differences in activation and deactivationAn ROI analysis was conducted to further examine the profile of

gender differences noted above in the between-group comparison.Weexamined differences in both activation (CalculationN Identification)and deactivation (IdentificationNCalculation) in the dorsal and ventralstream clusters identified in Fig. 3.

ROI analysis of the dorsal visual stream cluster revealed positivetask-related activation in males, whereas females did not activatesignificantly above the baseline (Fig. 4A). To further assess therobustness of these gender differences, we then examined the spatialextent of activation within these regions. To accomplish this, wecomputed the number of voxels in the cluster that were activatedabove a threshold of Z=|2.33|, (voxel-wise pb0.01). Males showedgreater activation than females, while females showed greater deac-tivation than males.

ROI analysis in the ventral stream cluster revealed a differentprofile: females showed significant deactivation, whereas males did

Fig. 4. Gender differences in dorsal and ventral visual stream activation during mental arithmpercentage of voxels deactivated in the (A) dorsal and (B) ventral visual streams. Gender d(IdentificationNCalculation) were observed. Height thresholds of Z=2.33 and Z=−2.33 (pmales showed significant (positive) activation whereas females did not activate significantgreater in females. In the ventral visual stream (B), females showed significant deactivation,greater in males and deactivation was greater in females.

not activate significantly above baseline (Fig. 4B). This was confirmedin a parallel analysis of the spatial extent of activation, with femalesdeactivating significantly more voxels than males.

Within the dorsal stream cluster, we further examined localizedresponses in the right IPS and the right angular gyrus using 10 mmspheres around the corresponding peaks of activation from thebetween-group comparison shown in Table 2. Both males and femalesshowed positive activation in the IPS, but males showed significantlygreater activation than females, p=0.007 (Fig. 5). In the right angulargyrus, however, males showed positive activation and females showeddeactivation. As a result, brain responses in the right angular gyruswere significantly greater in males (p=0.003).

We repeated the above analysis using 10 mm spheres around theactivation peaks within the ventral visual stream clusters (Table 2)encompassing the right lingual gyrus and the right PHG. Femalesshowed significant deactivation in both regions, while activation didnot significantly differ from baseline inmales (Fig. 6). As a result, brain

etic. (Left) Average signal strength, (center) percentage of voxels activated, and (right)ifferences in both task-related activation (CalculationN Identification) and deactivationb0.01) were used to assess activation and deactivation. In the dorsal visual stream (A),ly above baseline. Consequently, activation was greater in males and deactivation waswhereas males did not activate significantly above baseline. Accordingly, activation was

Fig. 5. Gender differences in the right posterior parietal cortex during mentalarithmetic. Signal strength in the right intra-parietal sulcus (IPS) (left), and the rightangular gyrus (right) in males and females. Significant gender differences wereobserved in both the IPS (p=0.007) and the angular gyrus (p=0.003). These genderdifferences were the result of both task-related activation (CalculationN Identification)and deactivation (IdentificationNCalculation). The right IPS showed activation in bothmales and females, while the right angular gyrus showed positive activation in malesand deactivation in females. 10 mm sphere ROIs about the peak activated voxels in theposterior parietal cortex within the dorsal stream (see Fig. 3) are shown here. The MNIcoordinates (x, y and z in mm) of the peaks are shown above each graph.

Fig. 6. Gender differences in the right lingual and parahippocampal gyri during mentalarithmetic. Signal strength in the right lingual gyrus (left) and in the rightparahippocampal gyrus (PHG) (right) in males and females. Significant genderdifferences were observed in the lingual gyrus (p=0.013) and the PHG (p=0.016).In both regions, gender differences were the result of significant task-relateddeactivation (IdentificationNCalculation) in females. Signal strength in males was notsignificantly different from baseline. 10mm sphere ROIs about the peak activated voxelswithin the ventral visual stream (see Fig. 3) are shown here. Other details as explainedin Fig. 5.

348 K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

responses were significantly greater in males, compared to females, inthe right lingual gyrus and the right PHG, p=0.013 and p=0.016,respectively.

Gender differences within the default mode networkBoth the dorsal and ventral stream clusters that differed in

functional responses between males and females overlapped withregions comprising the default mode network (DMN). The DMN is anetwork of brain regions that is consistently deactivated across a widerange of cognitive tasks (Greicius et al., 2003; Raichle et al., 2001). Asshown in Supplementary Fig. S1, overlap with the DMN was observedin the right angular and lingual gyri as well as the PHG, but not theright IPS. To further quantify brain responses in males and females inrelation to the DMN, we examined brain responses in the dorsal andventral stream clusters that overlapped with the DMN. This analysisrevealed that within the DMN, females tend to more stronglydeactivate both the dorsal and ventral streams (Supplementary Fig.S2, left hand column).

Outside the DMN, within the PPC regions of the dorsal stream,males showed significant positive activation, while females did notshow significant activation above baseline (Supplementary Fig. S2A,right hand column). Within the ventral stream, deactivation extendedoutside the DMN in females, whereas males showed positive task-related activation (Supplementary Fig. S2B, right hand column).

Gender differences in frontal lobe and subcortical structuresFrontal lobe and subcortical structures were extensively activated

in bothmales and females during the Calculation task, as seen in Fig. 2.To confirm the lack of gender differences observed in these regions,we used functional ROIs based on significant clusters of activationdetected in a previously published study of MA (Menon et al., 2000b).There was no overlap in participants of the present and prior studies.The left inferior and bilateral middle frontal gyri, right thalamus andleft cerebellum showed activation in males and females as measuredboth bymean signal strength (Supplementary Fig. S3) and percentageof voxels activated (Supplementary Fig. S4). No gender differenceswere seen in any of the frontal or subcortical regions for either averagesignal strength or spatial extent of activation.

Gender differences in brain–behavior relationsWe used a general linear model to examine gender differences in

the relationship between brain responses and behavioral measures(accuracy and RT). The analysis was conducted using both voxel-wisewhole brain analysis, as well as ROI analyses consisting of both thedorsal and ventral clusters, as well as the 10 mm sphere ROIs around

the peaks of activation in the IPS, angular gyrus, lingual gyrus, andPHG. In each case, there were no gender differences in the profile ofbrain–behavior relations as assessed using a difference of slopes test.

Gender differences in anatomyThe aim of this analysis was primarily to examine whether the

observed functional differences between males and females arosefrom underlying anatomical differences. VBM showed that femaleshad greater regional density and volume in a number of posteriorbrain areas where we had detected functional activation differencesbetween males and females (Fig. 7). Specifically, females showedgreater density bilaterally in the IPS, PHG, extrastriate cortex (BA 17),supramarginal gyrus (BA 40), angular gyrus (BA 39), superior parietallobule (BA 7), and inferior temporal cortex (BA 37), pb0.05, cluster-level corrected for multiple comparisons at pb0.05; Fig. 7, left column.Males, on the other hand, did not show greater density in any brainregion. Females also showed greater regional volume bilaterally in theIPS, supramarginal gyrus (BA 40), angular gyrus (BA 39), superiorparietal lobule (BA 7), and in the right extrastriate cortex (BA 17),pb0.05, cluster-level corrected for multiple comparisons at pb0.05;Fig. 7, right column. Again, males did not show greater regionalvolume in any brain region.

Discussion

Our study provides new insights into the organization of brainsystems involved in mathematical cognition in females and males.Although brain responses showed considerable overlap in manyregions that are known to be involved in MA, significant genderdifferences were also observed, despite the lack of overt behavioraldifferences. Gender differences were all localized to the right posteriorregions of the brain, where number and space interact (Ansari, 2008;Hubbard et al., 2005). Gender differences in activation were localizedto two clusters: one within the dorsal visuospatial stream, and theother within the ventral visuospatial processing stream (Goodale andWestwood, 2004; Milner and Goodale, 2007; Ungerleider andMishkin, 1982). Within the dorsal stream, males showed greateractivation than females in the right IPS and the right angular gyrusregions of the PPC; within the ventral visual stream, males showedgreater activation than females in the right lingual and the right PHG.Interestingly, therewere no brain areas where females showed greaterfunctional activation than males. In addition, significant differenceswere found in the structural density and volume of areas overlappingwith those activated by the mathematical tasks such that females hadgreater density and volume than males.

349K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

Similar levels of performance in males and females

To our knowledge, only two behavioral studies have used three-operand arithmetic problems similar to the ones used in our study.

Fig. 7.Gender differences in graymatter and its overlapwith functional differences. Left paneregions where females showed greater gray matter density than males. No brain areas showdifferences in gray matter density were observed in the two clusters encompassing the dorgreater functional responses than females. Brain regions that showed structural differences aoverlap in structural and functional differences are shown in purple. Right panel (A) showsshowed greater graymatter volume thanmales. No brain areas showed greater graymatter vovolume were observed in the dorsal visual stream regions where males showed greater fun

Our results are consistent with the findings from these studies inthat there were no differences between males and females in eitheraccuracy or RT (Geary et al., 2000; Royer et al., 1999). Geary et al.(2000) gave college-age participants 2 min to solve as many three-

l (A) shows unmodulated VBM results showing parietal, prefrontal and temporal corticaled greater gray matter density in males, compared to females. (B, C) Significant gendersal visual stream and ventral visual stream regions, respectively, where males showedre highlighted in cyan, functional differences are shown in red and the voxels that showmodulated VBM results showing parietal and prefrontal cortical regions where femaleslume inmales, compared to females. (B, C) Significant gender differences in graymatterctional responses than females.

350 K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

operand addition problems as possible, and observed no genderdifferences in the number of problems accurately solved during thetask (d=0.04). Although Royer et al. (1999) did not directlycompare performance between genders on their three-operandaddition task, the reported mean accuracies of 94.07 (SD=8.04)for males and 94.04 (SD=8.04) for females are not significantlydifferent. Similarly, reaction times did not appear to differ betweenthe genders, with a mean reaction time of 1.94 s (SD=0.783) formales, and 2.19 s (SD=0.801) for females. Our study extends thesefindings in two ways: (1) we used both addition and subtractionoperations in three-operand problems, and (2) we presented eachproblem in a time-limited format, giving participants only 4 s toanswer each problem. Our findings contribute to growing evidencethat gender differences in mathematics performance are small, evenin high achieving young adults. Taken together, these results providefurther support for the gender similarities hypothesis (Hyde, 2005),and argue against the notion of innate gender differences inmathematical calculation. Methodologically, these results are impor-tant here because the lack of gender differences in behavior allowsus to examine differences in the functional organization of brainareas independent of performance.

Gender differences in the dorsal visual stream

Within the dorsal stream, differences were observed in the rightIPS and the right posterior angular gyrus. The right IPS showedgreater activation in the MA, compared to the Identification task, inboth males and females (Fig. 5). However, the magnitude of thisresponse was significantly different—males showed greater activa-tion than females. The right IPS region where gender differenceswere observed is part of the set of canonical parietal lobe regionsthat are activated during MA tasks. Functional imaging studies haveshown that the IPS region plays a crucial role in arithmeticcalculations, independent of other processing demands such asworking memory (Delazer et al., 2003; Gruber et al., 2001; L. Zagoand Tzourio-Mazoyer, 2002). Behavioral studies in patients withdamage to the PPC demonstrate significant deficits in performanceof mental addition and subtraction tasks (Levin et al., 1993). Inaddition, the PPC has been classically thought to underlie theacalculia component of Gerstmann's syndrome (Levin et al., 1993),although the precise overlap with the IPS regions showingdifferences in our study remains unclear. No consistent view ofthe differential role of the left and right IPS in MA has yet emerged;but our findings suggest that males may engage the right IPS moreheavily to perform MA tasks. It is noteworthy that sex differences insimilar right PPC regions have also been reported in three-dimensional mental rotation tasks, with males showing greaterresponses (Hugdahl et al., 2006; Thomsen et al., 2000; Weiss et al.,2003). The precise foci of these gender differences differ acrossstudies, and the extent to which the two cognitive processesoverlap neurally is at present unclear. Nevertheless, these findingsare consistent with the view that numerical and visuospatialinformation processing overlap, at least partially, in the right IPS(Zago et al., 2008).

A different pattern of gender differences was detected in theright inferior PPC, in a region that includes the angular gyrus,extending posteriorly to the transition zone between the parietaland occipital cortex (Duvernoy et al., 1999). In this region, malesshowed slight activation, while females showed deactivation (Fig.5). One potential issue in interpreting these findings is that it leavesunclear whether the observed gender differences in deactivationarose from the activation of these regions during the Identificationcondition. In order to address this issue, we analyzed a differentfMRI dataset acquired in a separate group of 11 female and 10 maleparticipants with both number identification and passive fixation“rest” baseline conditions. We found no gender differences in the

number identification, compared to the passive fixation, conditionin the angular gyrus (Supplementary Fig. S5). Further, no genderdifferences were detected in either the dorsal or ventral streamclusters when we compared brain responses during the numberidentification condition with the passive fixation condition (Supple-mentary Fig. S6). These analyses indicate that gender differencesreported here arise from differences in deactivation during the MACalculation task as opposed to activation differences during theIdentification task.

Angular gyrus deactivation has been reported in several studies inresponse to greater cognitive load, and this region is also known tooverlap with the default mode network (DMN) (Greicius et al., 2003;Raichle et al., 2001). The DMN is a set of tightly coupled brain regionsthat consistently shows significant deactivation during demandingcognitive tasks. We found that angular gyrus regions that overlappedwith the canonical DMN (Greicius et al., 2003) showed significantdeactivation in females, with baseline activity in males (Fig. S4). Incontrast, angular gyrus regions outside the DMN showed significantpositive activation in males, with baseline activity in females.Although the functions of the lateral parietal lobe component of theDMN are not yet well understood, emerging evidence suggests thatangular gyrus regions noted above may play an important role inmemory retrieval (Vincent et al., 2006; Wagner et al., 2005). Takentogether, these findings suggest that the functional organization ofinferior parietal lobule systems that contribute to computational andretrieval processes during mathematical information processing isdifferent in males and females.

Gender differences in the ventral visual stream

Gender differences were also observed in the ventral visual stream.Like in the parietal cortex, these differences were restricted to theright hemisphere. Differences were observed in both the lingual gyrusand PHG, which form part of a hierarchically organized processingstream for encoding visual stimuli (Menon et al., 2000a) thatconverges into the hippocampus. Interestingly, no gender differenceswere observed in the hippocampus. Based on previous studies, thehippocampus does not play a critical role in the retrieval of MA facts inadults, since these facts are thought to be consolidated into long-termmemory, primarily in the PPC (Dehaene et al., 2003; Delazer et al.,2004; Rivera et al., 2005). Similar to the findings regarding the rightangular gyrus regions noted above, gender differences in the ventralvisual stream regions were also related to greater deactivation infemales (Fig. 6). The lingual gyrus and PHG regions are not generallythought of as being part of the canonical DMN, but these regionsvariably overlap with the posterior cingulate cortex regions which aretypically deactivated during high-level cognitive tasks (Greicius et al.,2003). Unlike the IPS and angular gyrus regions noted above, however,the specific role of these ventral visual regions in mathematicalinformation processing is less clear.

A similar pattern of gender differences in the ventral visual streamhas been reported during a mental rotation task in which the rightmedial lingual gyrus showed greater activation in males (Seurincket al., 2004). Interestingly, gender differences in the lingual gyrus havealso been observed during passive viewing of simple visual stimuli.For example, in one study, men activated the right lingual gyrus (BA18) 30% more than women in response to flickering checkerboardstimuli (Kaufmann et al., 2001). Unfortunately, these studies did notdetermine whether the gender difference in lingual activation wasprimarily due to activation by the males, or deactivation by thefemales, as was seen in the current study. Nevertheless, these studiesprovide converging evidence for gender differences in the functionalorganization of early visual pathways. The manner in which thesedifferences contribute to, or bias processing of, complex numericaland symbolic strings, such as those used in our MA task remains to beinvestigated.

351K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

Interpreting gender differences in brain activation

Our findings clearly indicate that although there was considerableoverlap in the brain areas activated by males and females during ourmathematical cognition task, there were significant differences inbrain responses as well. Our study includes one of the largest samplesizes for which gender differences have been examined in anyfunctional brain imaging study to date. The large sample size(N=49) and large effect sizes (dN1.1) suggest that the differencesobserved in the dorsal and ventral visual streams are likely to berobust and replicable. To our knowledge, this is the first study toreport effect sizes for gender differences in functional brain imagingstudies.When comparedwith small effect sizes reported in behavioralstudies, our results suggest that functional brain imaging may providemore sensitive measures of gender differences in informationprocessing than purely behavioral measures.

Our findings raise an intriguing question—how do we interpretdifferences in brain response in the absence of behavioral differences?One possible interpretation is that males and females might differ inthe cognitive strategy used, even though their performance levels aresimilar. In children, differences in visualization of multiple solutionpaths has been suggested as a possible mechanism of genderdifferences in strategy during math cognition tasks (Geary et al.,2000). To our knowledge, no such differences have been reported inadults. Given that our study contains well-educated participants, froma relatively gender-equal society (Guiso et al., 2008), it is unlikely thatmales and females applied different strategies to solve the fairlysimple MA tasks that we used here. In agreement with thisobservation, we did not find any gender differences between brain–behavioral relations in any brain region.

A more likely reason for the observed gender differences infunctional responses is that there are significant structural differencesbetween male and female brains in these regions. Our VBM analysisprovides strong evidence that this may indeed be the case. We foundthat females had greater gray matter density in both dorsal andventral stream regions where functional task-related gender differ-ences were observed. Specifically, within the dorsal stream, femalesshowed greater gray matter density in and around the right IPS andangular gyrus regions where males showed greater functionalresponses. Within the ventral stream, a similar pattern was observed,with females showing significantly greater gray matter density in andaround the right parahippocampal cortex regions where malesshowed greater functional activation. In addition, females showedgreater gray matter volume in dorsal stream regions where malesshowed greater functional responses. However, the ventral stream didnot show similar volumetric differences between the genders.

Interestingly, the structural differences observed in our study areconsistent with findings from several recent MRI and cytoarchitec-tonic studies. These studies found that females have thicker graymatter in temporal and parietal cortices after controlling for brain sizedifferences (Amunts et al., 2007; Im et al., 2006; Sowell et al., 2006).For example, Sowell et al. found that cortical thickness in the rightinferior parietal and posterior temporal regions was up to 0.45 mmthicker, on average, in females, even without correcting for thenormally smaller total brain volume in females (Sowell et al., 2006).These areas overlap with the right IPS region where we found genderdifferences in structure and function in our study. Furthermore,consistent with our finding of right lateralized gender differences inparietal lobe activation, structural imaging studies have revealedgender differences in hemispheric asymmetry within the parietalcortex and the parieto-occipital cortex gray matter, with males havinggreater asymmetry than females, especially in the right hemisphere(Amunts et al., 2007; Goldstein et al., 2001; Gur et al., 1999; Kovalevet al., 2003; Shah et al., 2004). Our study converges on the abovefindings and points to important gender differences in the functionaland structural organization of the right PPC. Exactly how these

structural differences could lead to functional differences in the extentof activation, and more surprisingly, in the magnitude of activationfound in the right IPS and angular gyrus is presently unclear. Wehypothesize that increased regional volume and density of graymatter in females may result in more efficient information processingmodules.

In agreement with this suggestion, a study of motor mentalrotation, where females and males performed equally well, found thatthe slope of the fMRI signal per degree of mental rotation was morethan two times higher in men than in women (Christova et al., 2008).In that study, such processing efficiencywas primarily restricted to theprefrontal cortex; weaker effects were also found in the right IPS,perhaps because of the small sample size used—four males and fourfemales. Our study, conducted with a much greater sample sizesuggests that similar processing efficiency principles may apply in thePPC in the context of mathematical cognition tasks. Taken together,these findings highlight the need for further studies to examine howfunctional modules are organized in males and females and how theydifferentially influence information processing.

Acknowledgments

We thank Sonia Crottaz-Herbette, Jason Hom and Caitlin Costellofor assistance with data acquisition and analysis, and Elena Rykhlevs-kaia for assistancewith the VBM analysis. This researchwas supportedby NIH grants HD047520 and HD059205, and NSF grant BCS/DRL0750340.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.neuroimage.2009.04.042.

References

Amunts, K., Armstrong, E., Malikovic, A., Homke, L., Mohlberg, H., Schleicher, A., et al.,2007. Gender-specific left-right asymmetries in human visual cortex. J. Neurosci.Official J. Soc. for Neurosci. 27 (6), 1356–1364.

Ansari, D., 2008. Effects of development and enculturation on number representation inthe brain. Nat. Rev. Neurosci. 9 (4), 278–291.

Armstrong, J.M., 1981. Achievement and participation of women in mathematics:results of two national surveys. J. Res. Math. Educ. 12, 356–372.

Ashburner, J., Friston, K.J., 2000. Voxel-based morphometry—the methods. NeuroImage11 (6 Pt 1), 805–821.

Ashburner, J., Friston, K.J., 2005. Unified segmentation. NeuroImage 26 (3), 839–851.Barbaresi, W., Katusic, S., Colligan, R., Weaver, A., Jacobsen, S., 2005. Math learning

disorder: incidence in a population-based birth cohort, 1976–82, Rochester. Minn.Ambul. Pediatr. 5 (5), 281–289.

Bell, E.C., Willson, M.C., Wilman, A.H., Dave, S., Silverstone, P.H., 2006. Males andfemales differ in brain activation during cognitive tasks. NeuroImage 30 (2),529–538.

Benbow, C.P., Lubinski, D., Shea, D.L., Eftekhari-Sanjani, H., 2000. Sex differences inmathematical reasoning ability at age 13: their status 20 years later. Psychol. Sci. J.Am. Psychol. Soc./APS 11 (6), 474–480.

Brett, M., (2000). The MNI brain and the Talairach atlas. http://www.mrc-cbu.cam.ac.uk/Imaging/mnispace.html.

Christova, P., Lewis, S., Tagaris, G., Uğurbil, K., Georgopoulos, A., 2008. A voxel-by-voxelparametric fMRI study of motor mental rotation: hemispheric specialization andgender differences in neural processing efficiency. Exp Brain Res. 189 (1), 79–90.

Dehaene, S., Spelke, E., Pinel, P., Stanescu, R., Tsivkin, S., 1999. Sources of mathematicalthinking: behavioral and brain-imaging evidence. Science 7 (284), 970–974.

Dehaene, S., Piazza, M., Pinel, P., Cohen, L., 2003. Three parietal circuits for numberprocessing. Cogn. Neuropsychol. 20 (3/4/5/6), 487–506.

Dehaene, S., Molko, N., Cohen, L., Wilson, A.J., 2004. Arithmetic and the brain. Curr. Opin.Neurobiol. 14, 218–224.

Delazer, M., Girelli, L., Grana, A., Domahs, F., 2003. Number processing and calculation-normative data from healthy adults. Clin. Neuropsychol. 17 (3), 331–350.

Delazer, M., Domahs, F., Lochy, A., Bartha, L., Brenneis, C., Trieb, T., 2004. The acquisitionof arithmetic knowledge—an FMRI study. Cortex; a journal devoted to the study ofthe nervous system and behavior 40 (1), 166–167.

Duvernoy, H.M., Bourgouin, P., Cabanis, E.A., Cattin, F., 1999. The Human Brain:Functional Anatomy, Vascularization and Serial Sections with MRI. Springer, NewYork.

Feingold, A., 1992. Sex difference in variability in intellectual abilities: a new look at anold controversy. Rev. Educ. Res. 62, 61–84.

352 K. Keller, V. Menon / NeuroImage 47 (2009) 342–352

Frings, L., Wagner, K., Unterrainer, J., Spreer, J., Halsband, U., Schulze-Bonhage, A., 2006.Gender-related differences in lateralization of hippocampal activation andcognitive strategy. NeuroReport 17 (4), 417–421.

Friston, K., Buechel, C., Fink, G., Morris, J., Rolls, E., Dolan, R., 1997. Psychophysiologicaland modulatory interactions in neuroimaging. NeuroImage 6 (3), 218–229.

Friston, K., Penny,W., Glaser, D., 2005. Conjunction revisited. NeuroImage 25 (3), 661–667.Gallagher, A.M., Kaufman, J.C., 2005. Gender Differences in Mathematics: An Integrative

Psychological Approach. Cambridge Univ. Press.Geary, D.C., 1996. Sexual selection and sex differences in mathematical abilities. Behav.

Brain Sci. 19, 229–284.Geary, D.C., 1998. Male, Female: The Evolution of Human Sex Differences. American

Psychological Association, Washington DC.Geary, D.C., Saults, S.J., Liu, F., Hoard, M.K., 2000. Sex differences in spatial cognition,

computational fluency, and arithmetical reasoning. J. Exp. Child Psychol. 77 (4),337–353.

Glover, G.H., Lai, S., 1998. Self-navigated spiral fMRI: interleaved versus single-shot.Magn. Res. Med.: Official J. Soc. Magn. Reson. Med./Soc. Magn. Reson. Med. 39 (3),361–368.

Goldin, C., Katz, L.F., Kuziemko, I., 2006. The homecoming of American college women:the reversal of the college gender gap. J. Econ. Perspect. 20 (4), 133–156.

Goldstein, J.M., Seidman, L.J., Horton, N.J., Makris, N., Kennedy, D.N., Caviness Jr, S., et al.,2001. Normal sexual dimorphism of the adult human brain assessed by in vivomagnetic resonance imaging. Cerebral cortex (New York, N.Y.: 1991) 11 (6),490–497.

Goldstein, J.M., Jerram, M., Poldrack, R., Anagnoson, R., Breiter, H.C., Makris, N., et al.,2005. Sex differences in prefrontal cortical brain activity during fMRI of auditoryverbal working memory. Neuropsychology 19 (4), 509–519.

Good, C., Johnsrude, I., Ashburner, J., Henson, R., Friston, K., Frackowiak, R., 2001.Cerebral asymmetry and the effects of sex and handedness on brain structure: avoxel-based morphometric analysis of 465 normal adult human brains. Neuro-Image 14 (3), 685–700.

Goodale, M., Westwood, D., 2004. An evolving view of duplex vision: separate butinteracting cortical pathways for perception and action. Curr. Opin. Neurobiol. 14(2), 203–211.

Greicius, M.D., Krasnow, B., Reiss, A.L., Menon, V., 2003. Functional connectivity in theresting brain: a network analysis of the default mode hypothesis. Proc. Natl. Acad.Sci. U. S. A. 100 (1), 253–258.

Gruber, O., Indefrey, P., Steinmetz, H., Kleinschmidt, A., 2001. Dissociating neuralcorrelates of cognitive components in mental calculation. Cereb. Cortex (New York,N.Y.: 1991) 11 (4), 350–359.

Guiso, L., Monte, F., Sapienza, P., Zingales, L., 2008. Diversity. Culture, gender, and math.Science 320 (5880), 1164–1165.

Gur, R.C., Turetsky, B.I., Matsui, M., Yan, M., Bilker, W., Hughett, P., et al., 1999. Sexdifferences in brain gray and white matter in healthy young adults: correlations withcognitive performance. J. Neurosci.: Official J. Soc. Neurosci. 19 (10), 4065–4072.

Halpern, D.F., Benbow, C.P., Geary, D.C., Gur, R.C., Hyde, J.S., Gernsbacher, M.A., 2007. Thescience of sex differences in science andmathematics. Psychol. Sci. Public Interest 8(1), 1–51.

Hubbard, E., Piazza, M., Pinel, P., Dehaene, S., 2005. Interactions between number andspace in parietal cortex. Nat. Rev., Neurosci. 6 (6), 435–448.

Hugdahl, K., Thomsen, T., Ersland, L., 2006. Sex differences in visuo-spatial processing:an fMRI study of mental rotation. Neuropsychologia 44 (9), 1575–1583.

Hyde, J.S., 2005. The gender similarities hypothesis. Am. Psychol. 60 (6), 581–592.Hyde, J.S., Fennema, E., Lamon, S.J., 1990. Gender differences in mathematics

performance: a meta-analysis. Psychol. Bull. 107 (2), 139–155.Ilg, R., Wohlschläger, A., Gaser, C., Liebau, Y., Dauner, R., Wöller, A., et al., 2008. Gray

matter increase induced by practice correlates with task-specific activation: acombined functional and morphometric magnetic resonance imaging study.J. Neurosci. 28 (16), 4210–4215.

Im, K., Lee, J.M., Lee, J., Shin, Y.W., Kim, I.Y., Kwon, J.S., et al., 2006. Gender differenceanalysis of cortical thickness in healthy young adults with surface-based methods.NeuroImage 31 (1), 31–38.

Jacobs, J.E., 2005. Twenty-five years of research on gender and ethnic differences inmath and science career choices: what have we learned? New Dir.Child Adolesc.Dev. 110 (110), 85–94.

Kaufmann, C., Elbel, G.K., Gossl, C., Putz, B., Auer, D.P., 2001. Frequency dependence andgender effects in visual cortical regions involved in temporal frequency dependentpattern processing. Hum. Brain Mapp. 14 (1), 28–38.

Kegel-Flom, P., Didion, C.J., 1995. Women, math, and test scores. Science 270 (5235),364–365.

Kim, D.H., Adalsteinsson, E., Glover, G.H., Spielman, D.M., 2002. Regularized higher-order in vivo shimming. Magn. Reson. Med.: Official J. Soc. Magn. Reson. Med./Soc.Magn. Reson. Med. 48 (4), 715–722.

Kovalev, V.A., Kruggel, F., von Cramon, D.Y., 2003. Gender and age effects in structuralbrain asymmetry as measured by MRI texture analysis. NeuroImage 19 (3),895–905.

Levin, H.S., Goldstein, F.C., Spiers, P.A., Heilman, K.M., Valenstein, E., 1993. Acalculia.Clinical Neuropsychology, Vol. 3. Oxford Univ. Press, New York, pp. 91–122.

Menon, V., Eliez, S., Glover, G.H., Reiss, A.L., 2000a. Analysis of distributed neural systeminvolved in spatial, novelty and memory processing. Hum. Brain Mapp. 11 (2),117–129.

Menon, V., Rivera, S.M.,White, C.D., Glover, G.H., Reiss, A.L., 2000b. Dissociating prefrontal andparietal cortex activation during arithmetic processing. NeuroImage 12 (4), 357–365.

Milner, A., Goodale, M., 2007. Two visual systems re-viewed. Neuropsychologia 46 (3),774–785.

Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L.,2001. A default mode of brain function. Proc. Natl. Acad. Sci. U. S. A. 98 (2),676–682.

Rivera, S.M., Reiss, A.L., Eckert, M.A., Menon, V., 2005. Developmental changes in mentalarithmetic: evidence for increased functional specialization in the left inferiorparietal cortex. Cereb. Cortex 15 (11), 1779–1790.

Royer, J.M., Tronsky, L.N., Chan, Y., Jackson, S.J., Marchant III, H., 1999. Math-fact retrievalas the cognitive mechanism underlying gender differences in math test perfor-mance. Contemp. Educ. Psychol. 24 (3), 181–266.

Seurinck, R., Vingerhoets, G., de Lange, F.P., Achten, E., 2004. Does egocentric mentalrotation elicit sex differences? NeuroImage 23 (4), 1440–1449.

Shah, N.M., Pisapia, D.J., Maniatis, S., Mendelsohn, M.M., Nemes, A., Axel, R., 2004.Visualizing sexual dimorphism in the brain. Neuron 43 (3), 313–319.

Sowell, E.R., Peterson, B.S., Kan, E., Woods, R.P., Yoshii, J., Bansal, R., et al., 2006.Sex differences in cortical thickness mapped in 176 healthy individualsbetween 7 and 87 years of age. Cereb. Cortex 17 (7), 1550–1560 (New York,N.Y.: 1991).

Stanley, J.C., Colangelo, N., Assouline, S.G., Ambroson, D.L., 1994. Gender differences forable elementary school students on above-grade-level ability and achievementtests. Talent Development: Vol. 2. Proceedings from the 1993 Henry B. and JocelynWallace National Research Symposium on Talent Development, Vol. 2. OhioPsychology Press, Vandalia, OH, pp. 141–148.

Thomsen, T., Hugdahl, K., Ersland, L., Barndon, R., Lundervold, A., Smievoll, A.I., et al.,2000. Functional magnetic resonance imaging (fMRI) study of sex differences in amental rotation task. Med. Sci. Monit.: Int. Med. J. Exp. Clin. Res. 6 (6), 1186–1196.

Ungerleider, L.G., Mishkin, M., 1982. Two cortical visual systems. In: Ingle, D.J.,Mansfield, R.J.W., Goodale, M.S. (Eds.), The Analysis of Visual Behaviour. MIT Press,Cambridge, MA, pp. 549–586.

Vincent, J.L., Snyder, A.Z., Fox, M.D., Shannon, B.J., Andrews, J.R., Raichle, M.E., et al.,2006. Coherent spontaneous activity identifies a hippocampal-parietal memorynetwork. J. Neurophysiol. 96 (6), 3517–3531.

Wagner, A.D., Shannon, B.J., Kahn, I., Buckner, R.L., 2005. Parietal lobe contributions toepisodic memory retrieval. Trends Cogn. Sci. 9 (9), 445–453.

Weiss, E., Siedentopf, C.M., Hofer, A., Deisenhammer, E.A., Hoptman, M.J., Kremser, C.,et al., 2003. Sex differences in brain activation pattern during a visuospatialcognitive task: a functional magnetic resonance imaging study in healthyvolunteers. Neurosci. Lett. 344 (3), 169–172.

Willingham, W.W., Cole, N.S., 1997. Gender and Fair Assessment. Lawrence ErlbaumAssociates, Hillsdale, NJ.

Zago, L., Tzourio-Mazoyer, N., 2002. Distinguishing visuospatial working memory andcomplex mental calculation areas within the parietal lobes. Neurosci. Lett. 331 (1),45–49.

Zago, L., Petit, L., Turbelin, M., Andersson, F., Vigneau, M., Tzourio-Mazoyer, N., 2008.How verbal and spatial manipulation networks contribute to calculation: an fMRIstudy. Neuropsychologia 46 (9), 2403–2414.